Multi-task Reinforcement Learning on Meta-World MT10 v1 (Fixed)
88Success RateMT-MH-SAC
Evaluation Results
| Method | Links | |
|---|---|---|
| MT-MH-SACImplementation=Reported in [43]2020.03 | 88 | |
| Soft Modularization (Shallow)Network Depth=Shallow2020.03 | 87 | |
| Soft Modularization (Deep)Network Depth=Deep2020.03 | 86.7 | |
| MT-MH-SACImplementation=Internal implementation baseline2020.03 | 85 | |
| MT-SACImplementation=Internal implementation baseline2020.03 | 44 | |
| Mix-ExpertImplementation=Internal implementation baseline2020.03 | 42.8 | |
| MT-SACImplementation=Reported in [43]2020.03 | 39.5 | |
| GLiBRL2025.12 | 25 | |
| Hard RoutingImplementation=Internal implementation baseline2020.03 | 20.8 | |
| SDVT2025.12 | 19 | |
| ECET2025.12 | 18 | |
| TrMRL2025.12 | 14 |